Ownership of contracts is a title question, and titles are straightforward to assign. Somebody in finance or procurement will say the word and mean it. Custody is a different question, and it cannot be answered from an org chart. Pick one vendor agreement, name it, and ask who can produce the executed copy today. Then pick a second one, at a different site, signed by someone who has since left. The first answer usually comes quickly. The second is where the exposure actually sits.
In most portfolios custody is distributed, and distributed unevenly. Some agreements are held at corporate. Some sit in a site office. Some are in a shared drive nobody maintains, or in the mailbox of a manager who left two years ago, or with outside counsel, or nowhere at all except with the vendor. None of those is the wrong place on its own. The difficulty is that the location varies by contract and cannot be predicted in advance, which means there is no reliable way to answer a question about the portfolio as a whole without going and looking, one agreement at a time.
The counterparty has no such problem
A vendor holds one copy of one agreement in one system. They drafted much of it, they benefit from it, and they have every reason to keep it somewhere they can find it in a minute. An operator holds hundreds of agreements across dozens of sites, accumulated through decades of staff turnover, in whatever format each era used. That asymmetry has nothing to do with diligence. It is a function of volume and time, and it runs in one direction only. When there is a disagreement about what was actually agreed, the party who can produce the document sets the terms of the conversation.
This is structural, not a performance failure
Contract portfolios accumulate rather than get built. Renewals execute at site level, often signed by whoever was running the building that year. Agreements arrive with acquisitions, transferring in force at closing without the institutional memory that explains why any of them were signed. Vendors are inherited from predecessors who documented informally, or did not document at all.
Nobody designed this. It is the residue of ordinary operating decisions, each made sensibly at the time, over a long enough period that no one person ever saw the whole. Which is why looking for the individual who dropped it does not work. There is no such individual, and the search costs you the cooperation of the people who could actually help you rebuild the record.
What it costs
World Commerce and Contracting, in research published with Ironclad in January 2026, put post-signature value loss at 11 per cent of contract value. That figure is cross-industry, so it should be read as directional rather than as a finding about Retirement Living and Long-Term Care specifically. What makes it useful is not the headline but the decomposition, because the causes are named and most of them run straight through custody:
- Missed savings from inadequate negotiation, 2 to 3 per cent
- Contract modifications that were never recorded, 2 to 3 per cent
- Renewal cost from insufficient forward planning, 2 to 3 per cent
- Untracked price escalations, missed obligations and the penalties that follow, 1 to 2 per cent each
Most of that list requires the document. A price escalation nobody can see is one nobody challenges, and a clause nobody can produce is one nobody enforces. None of it is misconduct. All of it is administration.
The arithmetic worth running is your own rather than the benchmark. Take annual vendor spend across the portfolio and apply the range. Then remember that avoidable loss of this kind comes off operating income rather than off revenue, so at a 30 per cent margin every dollar of it takes more than three dollars of new revenue to replace. That is the number that makes a registry look inexpensive.
Where it surfaces
It rarely announces itself as a contract problem. It surfaces in diligence, when a lender or an acquirer asks for a complete schedule of agreements in force at a site, with term end dates and assignment provisions, and the schedule has to be assembled rather than exported. Assembling it takes weeks. It involves calling vendors to ask what was agreed, which tells the vendor something you would rather they did not know. And it produces a document that the person compiling it will not want to certify.
A test that takes an hour
Which of your major contracts expire in the next 90 days?
Answering that requires the paper. Not a platform and not a project, just the underlying documents somewhere they can be read. If producing the answer takes more than an hour, the constraint is not contract management. It is that no single record exists, and everything downstream depends on one: renewal decisions, benchmarking, obligation tracking, and the schedule a lender will eventually ask for.
The fix is not a stricter policy or a better spreadsheet. It is one registry that holds the executed document, the term dates, the notice windows and a named owner, maintained as a condition of operating rather than as a project with an end date. An operator who has that answers the 90-day question in a minute. An operator who does not will keep answering it in weeks, and will keep finding out what was agreed at the least convenient moment available.
Solving the critical path.
Source: World Commerce & Contracting with Ironclad, Closing the Procurement Value Gap, January 2026.